What Knowledge is Found in White Space?

“White space” is the term printers use to describe the empty space on a page. This includes the space between words, blank lines between paragraphs, the margins at the edges of the page, and blank pages at the end of chapters or at the beginning or end of a book. Is it nothing? Is it wasted space? Well, it’s actually not “nothing,” because if you took the white space all away you’d find it very difficult to read. Imagine a newspaper with no space between the columns. Or imagine black and white pictures printed on black paper. 

Removing the white space leads to a loss of information, and yet, questions still remain: 

Does that prove the white space contains information? 

How can we decipher and interpret that information? 

Is it worth the effort? 

Some might say this is just semantics. The loss of information is not because we removed the white space, but because we introduced noise. If we were listening to a concert and someone introduced other sounds in between the notes, it would result in audio nonsense as chaotic as our newspaper lacking column gaps. The analogy doesn’t fully fit; however, and it would be fairer to imagine we removed the timing between the notes. Now our concert will still sound like something, just not what we expected when we looked at the program. Let’s accept that there is minimal information that a tabula rasa provides, which is the invitation to provide more information.This seems only barely more than a semantic distinction. 

At the other end of the spectrum is another semantic issue I’d like to set aside. The extreme example is the blank page as a message. I’ll let you think about that for a bit.

Somewhere in here is the issue of “medium as the message.” The conveyance of our message communicates something. In education, the obvious example is what I have talked about previously―some knowledge is better conveyed visually. That is true for video and auditory knowledge as well. Rene' Magritte was famous for these types of messages. Magritte’s messages were very thought provoking though rather than educational.

Is a Table a Concept Map? 

For those used to working with concept maps, a table is nothing like a concept map. For those only partly familiar with concept maps, a table is easier to create and reap a benefit. For those unfamiliar with concept maps, they are worth learning about. It is literally a map of how concepts or facts are related to each other. It has a lot in common with “semantic networks”.  Here's an example:

For example, we could place the idea of “extensors” and “flexors” on our map and then place the concepts “brachioradialis”, “triceps”, “iliopsoas”, and “hamstrings” and start connecting all these terms in a way that indicates which muscles are extensors and which are flexors. We could add “upper arm” and “lower arm” and add more connections. We could wander off into “decerebrate posturing” and “decorticate posturing” and to many other concepts besides and soon have a mess of interconnections that should remind us of the picture printed on black paper mentioned above. This is not a useful concept map, but does approach the function of a semantic map. 

The vast number of interconnections might overwhelm our perception, but would only present a quantitative problem to a computer. In a sense, a semantic network provides a computable definition of the information it contains.The computer would be able to answer both the question “What is the triceps muscle?” and “What is the major extensor muscle of the upper arm?” There is some evidence that our brains are organized along the lines of semantic networks with specific areas for facts (nouns) and relations (verbs).

I see a continuum between well-written prose, the organization of a well-designed table, a well-designed concept map, and a semantic network of computable definitions. 

The semantic networks are leveraged with machine learning algorithms for things like autocompletion and optimizing search results. If we imagine adding some white space to them, or pruning them, they can help our less algorithmic-minded brains. Tables start from the other direction. They are easy to construct because we limit ourselves to a relatively small number of relationships. If we are lucky in picking the right initial relationships, then it feels like our table builds itself. The same thing that makes them easy also makes them constrained and limiting. We start to discover all the exceptions to the partial patterns that don’t quite fit in the tables. 

Concept maps fall somewhere in between. We don’t try to capture all the facts and relationships, just the most important ones. What many proponents of concept maps will tell you is that as you build them, you will discover the organizing principles and the more you build concept maps, the better you will get at it. This is poor advice for the anxious student who wants to believe this will help them, but only feels they are spending more time than they normally would on a topic and not making much progress. 

How Do People Create or Find “Organizing Principles?”

It is true that looking at a collection of similar and related information, people can be very good at seeing patterns. We generalize and make distinctions instinctively. A walk through a parking lot will grab our attention when we see a motorcycle, a three-wheeled car, or a sub-sub compact car―unless we live somewhere where those things are frequent. If someone is asked what you would expect to find in a parking lot, most people would answer cars. On the other hand, if we asked people to list all the things besides cars you might find in a parking lot, a fairly long list could be generated. In the absence of a specific goal, our brain will still attempt to provide an organizing principle to simplify what we are seeing, if only so it can ignore the generalization as white space to be ignored. 

Try the following exercise. Using just your imagination, visualize walking through a parking lot. Imagine you are looking for “your car”. I’ll let you think about that for a bit. 

Having done the exercise, now answer the question, “What was the predominant color of the cars in the imaginary parking lot?” The responses people give reveal some interesting things. Some people will have no special color in mind while others will. And still others will impose a default color, but only after the question has been asked (sort of like trying to remember a dream on waking, and then wondering if that happened in the dream, or you only supplied that explanation after waking.) Black, gray, white, blue, and red are common colors. If it was an average of all cars we’ve seen we’d expect the color to be brown but that’s not what most people report. 

It is also true that repeated attempts to find patterns in the things around us will improve and diversify our ability to generalize and categorize. However, there are steps to get better at this faster. Now do the same exercise except instead of looking for your car, look for the most unusual car in the imaginary parking lot. How do the colors in this parking lot compare to the previous parking lot? Likely you “found” an unusual car only to realize there was an even more unusual car to be found. You likely found yourself in a parking lot with more unusual than mundane cars. If it had been a real parking lot, the cars would be universally more mundane but you would be likely to notice subtle but unique features you wouldn’t notice if you were just looking for “your car”. 

Step 1: Decide what purpose this map is supposed to serve

What this exercise suggests is that the first organizing principle for creating a table or a concept map is to decide what purpose this map is supposed to serve. What questions might it need to answer? You can’t know for sure the perfect question to focus on but consider these possibilities: 

How will this information help me be a better physician?

How will this information help me understand recurring principles of basic science?

How will this information help me pass the test?

Step 2: Start Adding Concepts and Connection

With a focus in mind, the second step is to start adding concepts and connections. A concept map allows these to be anything while a table will need you to decide early on what the row and column organizing principles should be. Having chosen these principles for a table, extending the table becomes easy and usually has a clear end. With a concept map, adding new information in different ways is easy but it can be unclear what the endpoint is.

Step 3: Start Looking for Generalizations and Outliers

The third step is to start looking for generalizations and outliers. Can the common elements be grouped to reduce complexity? Are the outliers important exceptions or just tangential elements that can be removed and placed in some future table or location? In a table this can mean repositioning entire rows or columns to bring similar or contracting information into conjunction or imposing a useful order so finding information is easier (e.g. alphabetic or numeric.)

Step 4: Step Back and Rethink What You Created

The final step is to take a step back and rethink what you created. Are some parts too detailed suggesting maybe they need to be simplified? Are parts of the information unclear or confusing? Does the lack of clarity suggest some better organizing principles? If you are creating your tables in a spreadsheet to begin with it can make reorganizing, trimming, and grouping can be less onerous. Concepts maps are all about reorganizing and probably the part newcomers dislike the most. There is software to make different aspects of this less onerous as well so it might be worth trying a few out before settling on one.

How White Space Improves Understanding

But really, what would an article about white space be without an example that shows how white space improves understanding?

Organizing principles for efficient knowledge representation

• What is the purpose it will serve?
• What are the broad groupings that capture most information?
• Is there a level of detail that would be useful for general, intermediate, and expert levels? What information drops out at the expert levels? 
• What is the less essential information that can be ignored to serve the purpose?
• What is the optimal amount of interconnections that reveal interesting connections without being overwhelming? 
• How many levels are useful in your hierarchy and can greater delta be spun off into separate organizing efforts?

Whitespace is really about directing focus to specific material. The skillful use of whitespace is an essential tool when confronting the massive amounts of information such as we deal with in medical education. It is the invisible knowledge that keeps us from drowning from information overload.

View more of our recent blog posts.